GIACIATION OF THE YAKUTAT BAY REGION 227
evidence of the nature of the phenomena that result from such overriding, applicable to other regions.
Vegetation in Glacial Deposits. The observed growth of vegetation at the margin of glaciers, and even on them, and its incorporation in marginal deposits during an advance, point a clear warning against too free use of this class of evidence in the interpretation of the complexity of ancient glaciation. The phenomena of Yakutat Bay glaciation prove conclusively that soil beds, peat beds, and forest trees do not of themselves contribute evidence of interglacial conditions. They may represent no more than a half century or century of stagnation followed by moderate advance. This does not disprove interglacial conditions, nor argue against complexity of the Glacial Period, but it does discredit some of the evidence upon which some of the inferred complexity is based. More than incorporated vegetation is necessary for proof of such complexity.
Origin of Eskers. Few eskers were observed in the Yakutat Bay region, and these were short ones on overridden gravels in Russell Fiord. The absence of eskers is believed to be due in the main to their burial beneath the fringing alluvial fans whose apexes are at the ice tunnel from which the esker-building stream emerges, and which follow up the ice as it recedes. Esker-building conditions apparently exist in the piedmont bulbs, for the streams maintain their courses for a long time.
Our observations prove three important points with respect to eskers (PI. LXXXIX, A). (1) As already stated, that subsequent alluvial fan deposit may bury previously made eskers. (2) That, even under such favorable conditions as exist in the broad, stagnant, moraine-covered glaciers of this region, eskers are not formed on the glacier surface, for there are no long streams. (8) That moving glaciers are unfavorable to esker formation. The advance of Marvine and Variegated Glaciers completely destroyed pre-existing subglacial drainage; and the point of emergence of glacial streams from such a normally-moving glacier as the Nunatak is frequently changed. From the evidence of the Yakutat Bay region we believe that the eskers are subglacial deposits mainly, if not entirely, associated with stagnant ice conditions.
Outwash Gravels. The extensive sheets of gravel and the rapidity of their deposit in such valleys as that of Hidden Glacier, the absence of growth of vegetation on the out-wash gravel plains, and the perfect rounding of the pebbles, are phenomena which throw light on the extent and depth of similar deposits in regions of former glaciation. From a study of such an outwash gravel plain as that which existed in the Hidden Glacier valley in 1899, 1905, and 1906, or that of the Kwik River valley, we can observe the processes of accumulation which undoubtedly existed in thousands of streams issuing from the ancient continental glaciers.
Pitted Plains. Where these gravels rest on buried ice, as in the Hidden and Fourth Glacier valleys, the phenomenon of kettle formation and the production of pitted plains is observed in progress of development, and an explanation of thousands of similar cases is found.
Kames. Where gravels resting on ice settle very irregularly, we see in progress one of the important causes for the formation of kames; and this process is again and again illustrated in the Yakutat Bay region.
Buried Ice Blocks, Perhaps the most important contribution to an interpretation of glacial phenomena in regions of former glaciation yielded by the Yakutat Bay region is the clear evidence of the great importance of stagnant buried ice in shaping the topog-